Wallart commented on issue #16754: Is mirroring working with MXNet 1.5.1 Gluon ?
URL: https://github.com/apache/incubator-mxnet/issues/16754#issuecomment-561608565
To test mirroring, I've built a quick example, however after the first iteration the loss
is always nan. I cannot push investigation further.
I am missing something but I can't find what it is. I suppose that the Transformer LayerNorm
/ Dropout are not working in training mode, but using an autograd.train_mode() when I'm extracting
the network symbols does not solve the issue.
https://gist.github.com/Wallart/c807af4fdd520a172f41b7e17badc296
Finding good examples on Module API with the 404 errors on MXNet documentation feels like
archeology :D
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
With regards,
Apache Git Services
|